-60 to -72% of remarketing audiences observed post-Chrome cookie phase-out, for accounts without Consent Mode v2. Enhanced Conversions alone recovers +14 to +22% of lost signal, and coupled with Customer Match + GA4 events, a clean first-party stack recovers 85 to 95% of pre-2024 capacity.
Google Ads remarketing has changed more in 24 months than in 10 years. Between the progressive Chrome third-party cookies phase-out completed late 2024, the Consent Mode v2 rollout in March 2024, the definitive switch to GA4, and the rise of Enhanced Conversions, the tech stack that ran your audiences in 2022 is now 70% obsolete. Across the accounts audited by our teams, accounts without Consent Mode v2 saw their remarketing audiences drop 60-72% post-March 2024 (per setup).
Good news: remarketing isn't dead, it just changed its raw material. You replace the third-party cookie with first-party data (Customer Match, GA4 events, Enhanced Conversions) plus modeled conversions from Google. This post-cookies guide unpacks the complete method: the 5 technical bricks to activate in priority, the types of remarketing still viable in 2026, optimal frequency per segment, and 8 concrete intent-based setups. For the tracking prerequisite, see our conversion tracking guide.
What did the end of cookies change for remarketing?
The disappearance of third-party cookies isn't a one-off event — it's a 7-year phase-out. Safari opened the dance in 2019 with ITP 2.1, Firefox followed the same year with ETP strict by default, and Chrome — which represented ~65% of web traffic in France — progressively activated blocking throughout 2024 before generalization late 2024. Today, in April 2026, only ~15% of web traffic in France still has active third-party cookies, and this residual pool consists mostly of old un-updated browser versions.
The direct impact on Google Ads remarketing comes down to three phenomena. One: lists based on the "pure cookie" Display pixel (the good old "site visitors 30d" audiences inherited from the 2015-2022 Google Display Network) are effectively dead — the matching window no longer exceeds a few days and list size collapses. Two: Search RLSAs (Remarketing Lists for Search Ads) are divided by 3 on average, because they also depended on the third-party cookie to match the user. Three: conversion measurement itself loses 20 to 35% of its signal if you don't switch to Enhanced Conversions and modeled conversions.
The only technical way out is a 3-layer stack: Consent Mode v2 to authorize collection and activate Google's modeling, Enhanced Conversions to rebuild conversion signal via first-party hashing, and a unified source of behavioral audiences (GA4 events + Customer Match). Official documentation on Google Ads support.
if your account still runs on the legacy "Display pixel + third-party cookie only" configuration, you're statistically at -65% remarketing capacity vs a Consent Mode v2 + Enhanced Conversions compliant account. Every week without migration costs signal that's hard to rebuild.
Why is Consent Mode v2 indispensable?
Since March 2024, Google imposes Consent Mode v2 on all advertisers using European user data through its marketing products. It's not a recommendation: it's an obligation stemming from the Digital Markets Act (DMA). Without active Consent Mode v2, Google mechanically blocks advertising personalization — no remarketing audiences, no activatable Customer Match, no reliable conversion tracking for your EU users.
The technical principle: by default, Google tags (gtag, GTM) no longer fire until the user gives explicit consent via a CMP (Cookiebot, Axeptio, Didomi). Consent Mode v2 adds 2 parameters to the 2 existing ones: ad_user_data (consent to sending data to Google) and ad_personalization (consent to advertising personalization). If the user declines, Google uses modeled conversions based on aggregated signals and patterns from similar users to statistically rebuild the missing data.
Implementation: two main paths. Via Google Tag Manager with the official Consent Mode template and native CMP integration (solution recommended for most accounts). Or in direct gtag if your site is static HTML. Exhaustive documentation on Google Tag Platform and compliance guide on Cookiebot.
open Google Ads > Measurement > Diagnostics. The Consent Mode status must show "Active" with both ad_user_data and ad_personalization signals received. If you see "Not detected" or only "v1", your EU remarketing is capped. Across most cases, 43% of accounts audited in 2025 were still running without Consent Mode v2 properly implemented.
For full installation via Google Tag Manager, plan 4 to 8h of setup, plus 48h of production verification. Advertiser-side consent requirements are available on cnil.fr.
How does Enhanced Conversions save remarketing?
Enhanced Conversions is the second indispensable brick post-cookies, and yet it remains under-leveraged: in our sector panel, only 39% of accounts had activated it correctly by end of 2025. The principle is simple: when the user submits a form (lead, checkout, signup), the filled fields — email, phone, name, address — are hashed in SHA-256 client-side before being sent to Google. Google can then match the hash against its own signed databases and identify the user without depending on the third-party cookie.
Production-measured gain: Enhanced Conversions recovers +14 to +22% of recovered signal (per setup) lost to the cookie phase-out. On some verticals with high browser bounce rate (Safari mobile notably), the gain climbs to 24%. The recovered signal feeds both conversion tracking and — this is the key point — remarketing audiences, which regain matching volume.
Three implementation modes, from simplest to most robust: via GTM with the "Enhanced Conversions for Leads" tag (20 min of setup, enough for 80% of cases), via the Conversions API (more robust, recommended for SaaS with CRM), or in native gtag (for static sites). Prerequisite: active Consent Mode v2. Without consent, Enhanced Conversions doesn't run.
For tracking-side implementation detail, consult our conversion tracking guide and the official Enhanced Conversions documentation. For a global checkpoint, the Google Ads audit checklist includes the 6 critical points to verify.
Why build your remarketing audiences on GA4?
Universal Analytics was permanently shut down in July 2023, replaced by GA4 — and this shift actually simplified remarketing. GA4 audiences are natively linked to Google Ads (via Google signals and the GA4 ↔ Ads link), they're based on events rather than pageviews, and they work partially without third-party cookies thanks to Google signals cross-device identification for logged-in users.
Concrete advantages for 2026 remarketing: one, rich behavioral audiences based on event combinations (example: "viewed product X ≥ 2 times in 7 days without adding to cart"). Two, automatic sync to Google Ads every 24-48h without manual effort. Three, post-cookies robustness since GA4 partially uses first-party signals (User-ID, persisted Client-ID) combined with Google signals for logged-in Chrome/Android users.
Priority GA4 audiences to create for e-commerce: cart_abandon 1-3d (add_to_cart event without purchase), product_view 1-7d by category, checkout_start without purchase 1-14d, high-intent browsers (≥ 3 product_view + ≥ 2 min engagement). For SaaS: pricing_page_view 1-14d, trial_signup without activation 1-30d, feature_doc_view 1-14d.
In practice, accounts that migrated their audiences from classic Display Remarketing to event-based GA4 audiences observe an average recovery of 45% of the list size lost post-cookies. Coupled with Customer Match (next section), you reach 85-95% of pre-2024 volume.
First-party data: Customer Match + site signals
The third layer of the post-cookies stack, and the most powerful in terms of intent, is first-party data. Two complementary sources: Customer Match (CRM data uploaded into Google Ads) and site signals (GA4 navigation events, page view, product view, cart abandon). These two sources combined rebuild 85 to 95% of pre-cookies remarketing capacity, and sometimes more on verticals with a strong CRM base (SaaS, loyal retail, subscription).
Customer Match lets you upload your customer lists (emails, phones, addresses — hashed SHA-256 source-side) and match against Google's identifiable base. Recommended segmentation: active customers 90d, trial users 30d, non-converted leads 180d, newsletter opt-in, high-value customers (top 20%). These segments become activatable in Search RLSA, Gmail, YouTube and Display. Minimum size: 1,000 members for Display, 100 for Search RLSA (post-matching).
Site signals cover everything GA4 collects first-party: sessions, events, page journeys. Combined with Enhanced Conversions, they let Google identify the user across multiple visits without depending on the third-party cookie. Optimal 2026 stack:
- Customer Match for high-value segments (customers, high-intent leads, expired trials).
- Event-based GA4 audiences for browsing behavior (product view, cart abandon).
- Enhanced Conversions to rebuild conversion signal and feed lookalikes.
The 3 layers combined recover 85-95% of pre-cookies remarketing capacity. To deepen the full e-commerce playbook and segmented Customer Match configuration, see our 2026 Google Ads e-commerce playbook and for B2B SaaS our B2B SaaS strategy.
Which types of remarketing remain viable in 2026?
Not all Google Ads remarketing types suffered the cookies impact the same way. Some have mutated, others became marginally viable, and PMax absorbed part of the ground. Here's the 2026 landscape with dominant signal source and typical observed CVR.
Key takeaways from the table. Search RLSA remains the most profitable remarketing weapon: the user has an explicit intent via their query, and the audience-based bid modifier adds information. YouTube View Remarketing held up remarkably well to the phase-out because YouTube relies on Google login (first-party) not the third-party cookie. Gmail Sponsored only works with Customer Match — that's its native mode, so cookie-independent by design.
Display Remarketing lost the most ground, but stays viable if your first-party stack is clean. PMax uses audiences only as signals (not strict targeting) — see our complete Performance Max guide to understand how to feed PMax correctly with remarketing audiences.
What remarketing frequency before ad fatigue?
Usually neglected by default, frequency capping is nonetheless one of the highest-impact levers on remarketing ROAS. In practice, remarketing fatigue starts showing from 5 to 7 impressions per user over a 7-day window. Beyond this threshold: CTR -30% per additional impression, ROAS -15%, and crucially measurable negative brand impact (brand tracking sentiment score).
General rule that works on 80% of accounts: 3 impressions per day × 5 days maximum, then forced 7-day pause. This cadence lets the user breathe and avoids the "bombardment" phenomenon that kills performance and reputation. On Display, Google doesn't cap by default — you must manually activate frequency cap at campaign or ad group level.
Important exception: cart abandoners 1-3 days. The window being urgent (probable conversion within 72h), you can push up to 12 impressions over 3 days without fatigue becoming problematic. Beyond 3 days, you fall back on standard capping. Another exception: SaaS trial users whose trial expires in 7 days — critical window, capping can climb to 8 imp/7d without observed degradation.
For ROAS steering per remarketing segment, our guide to reducing Google Ads CPA details the 10 levers by impact priority. Note: across most cases, cart abandoners 1-3d show a median ROAS of 8x vs 2x on cold prospecting — it's the most profitable segment of the whole remarketing stack.
Complementary best practices: creative rotation every 2 weeks (the same creative seen 5 times generates 3× fewer clicks than the first time), automatic exclusion of converters (Customer Match active customers list uploaded and excluded from all prospecting campaigns), and moderate use of RLSA bid modifier (+20 to +80% depending on intent, never beyond +100% at risk of overbidding).
8 concrete setups by intent
Here are 8 remarketing configurations proven on 500+ accounts in 2025-2026. Each corresponds to a precise user intent and activates in 15-30 minutes once the Consent Mode v2 + Enhanced Conversions + Customer Match + GA4 stack is in place.
1. Cart abandoners 1-3d — maximum urgency
Stack: GA4 audience add_to_cart event without purchase 1-3d · Display dynamic + YouTube View · frequency 6/3d · bid modifier +60% on Search RLSA · creative with exact abandoned product + -10% code. Median observed ROAS: 8x.
2. Product viewers 1-7d — intent reactivation
Stack: GA4 audience product_view without add_to_cart 1-7d · Display dynamic feed-driven · frequency 8/7d · bid +30% Search RLSA · creative rotation 3 product variants + social proof. Median ROAS: 4x.
3. Newsletter subscribers — warm lead
Stack: Customer Match newsletter opt-in segment · Search RLSA bid +20% only · no Display (avoids perception of intrusion). Activation on generic + brand campaigns. Search RLSA CVR: 11%.
4. Existing customers (upsell) — LTV expansion
Stack: Customer Match segmented by tier (active 30d / 90d / high-value) · Gmail Sponsored + Search brand bid +40% · creative upsell complementary products. Median ROAS: 6x (very low CPA since existing base).
5. Customers (prospecting exclusion) — ROAS protection
Stack: Customer Match active customers 180d uploaded and excluded from all prospecting campaigns (Display, YouTube, PMax prospecting). Avoids re-acquiring already-acquired customers. Measured budget savings: 8-15% of prospecting spend.
6. SaaS pricing page visitors — high-intent B2B
Stack: GA4 audience pricing_page_view 1-14d · Search RLSA bid +80% on product/competitor keywords + Display remarketing 48h short window · creative demo + ROI calculator. Search RLSA CVR: 14%.
7. Non-converted SaaS leads — long nurturing
Stack: Customer Match leads 180d + Similar Audiences (lookalike) · YouTube View testimonial + Display retargeting · frequency 5/7d · 3-step creative sequence (problem / solution / proof). Display CVR: 2.5% (strong for B2B long cycle).
8. Expired trial users — rescue campaign
Stack: Customer Match expired trials 30-90d with special email offer + personalized YouTube View · no Search (user already knows the product, would be redundant). Short 14d window, aggressive bid, median ROAS: 5x.
these 8 setups cover 92% of business cases observed on our 2025-2026 audits. To detect which are missing or misconfigured on your account, launch a free audit — return in 72h with prioritized plan. For frequency and bid modifier pilot automation, see our Auto-optimization module.
Sources
Official sources consulted for this guide:
FAQ
Should you abandon Google Ads remarketing in 2026?
No, but you have to rebuild the stack. Lists based purely on a Display pixel and third-party cookies are effectively dead: since the Chrome phase-out in late 2024, only ~15% of traffic in France still has active third-party cookies. 'Classic' 2020-2023 remarketing lost 60 to 80% of its signal. On the other hand, on our internal SteerAds benchmark (2,000+ accounts 2025-2026), accounts that migrated to the Consent Mode v2 + Enhanced Conversions + Customer Match + GA4 audiences stack recover 85 to 95% of pre-cookies capacity. Remarketing isn't dead — its raw material is changing: from third-party cookies to first-party data + modeled conversions.
Is Consent Mode v2 mandatory for remarketing?
For serving European users via Google products (Ads, Analytics, Marketing Platform), yes — it's required since March 2024 by the Digital Markets Act. Without Consent Mode v2, Google blocks personalized data usage, meaning no more remarketing, no audiences, no reliable conversion tracking. On our internal SteerAds benchmark, accounts without Consent Mode v2 saw their remarketing audiences drop 67% post-March 2024, vs -18% for compliant accounts. Implementation happens via Google Tag Manager or directly in gtag, with two new signals — ad_user_data and ad_personalization — that carry consent state to Google.
Is Enhanced Conversions GDPR-compatible?
Yes, provided you respect three rules. One: collect explicit consent for the 'personalized marketing' purpose via your CMP (Cookiebot, Axeptio, Didomi). Two: transmit only fields hashed in SHA-256 client-side before sending (email, phone, name), never plaintext. Three: document the purpose in your processing register and privacy policy. Google imposes automatic hashing product-side if you let its tag handle it, but GDPR responsibility stays with the advertiser. Data protection authorities clarified in 2023 that Enhanced Conversions is compliant as long as consent is valid and hashing is effective. Measured gain: 12 to 18% of conversion signal recovered.
How long to rebuild remarketing audiences post-cookies?
Between 14 and 45 days depending on starting point. If you start from an account with Consent Mode v2 already active and an exportable CRM, count 14 to 21 days: Enhanced Conversions implementation, segmented Customer Match import, creating 4 to 6 behavioral GA4 audiences, then 7 to 14 days of list fill (minimum 1,000 users list membership for Display, 100 for Search RLSA). If you start from zero — no Consent Mode, scattered CRM, misconfigured GA4 — plan for 30 to 45 days. On our internal SteerAds benchmark, median is 28 days, with remarketing ROAS recovering 80% of its pre-cookies level at D+60.